An efficient Dai-Kou-type method with image de-blurring application

Iranian Journal of Numerical Analysis and Optimization

مقالات آماده انتشار، پذیرفته شده، انتشار آنلاین از تاریخ 11 خرداد 1404

نوع مقاله: Research Article

شناسه دیجیتال (DOI): 10.22067/ijnao.2025.92708.1615

نویسندگان

Kabiru Ahmed^{* 1}؛ Mohammed Yusuf Waziri¹؛ Salisu Murtala²؛ Abubakar Sani Halilu³؛ Habibu Abdullahi³؛ Ya'u Balarabe Musa³

¹Department of Mathematical Sciences, Bayero University, Kano, Nigeria.

²Department of Mathematics, Federal University, Dutse, Nigeria.

³Department of Mathematics, Sule Lamido University, Kafin Hausa, Nigeria.

چکیده

Well-conditioning of matrices has been shown to improve the numerical performance of algorithms by way of ensuring their numerical stability. In this paper, a modified Dai-Koutype conjugate gradient method is developed for constrained nonlinear monotone systems by employing the well-conditioning approach. The new method ensures that the much required condition for global convergence of iterates generated is satisfied irrespective of the linesearch strategy employed. Another novelty of the scheme is its practical application in image de-blurring problems. The method performs well and converges globally under mild assumptions. Experiments in image de-blurring and convex constrained system of equations, show the scheme to be effective.

کلیدواژه‌ها

Nonlinear equations؛ Eigenvalues؛ Constrained Equations؛ Convex Set؛ Sparse Signals